Out of Site

March 27, 2008 1:29 PM

Why does one particular website refuse to load?

I try to connect to: sbc-services.com but get a 'server not responding' message.

I've tried pinging it using terminal... nothing. But, if I try pinging it from a remote computer, everything seems fine.

I've switched everything on and off again but the problem persists. My girlfriend connects through the same router (both of us use wi-fi via an airport express thingy) and also has the same problem.

I changed the router's DNS settings from those of my ISP, to those of OpenDNS ...problem still there.

Huh? What do I do? I'm tired.

posted by popcassady to Computers & Internet (29 answers total)

What browser are you using? I just loaded it fine in Firefox, if that helps.
posted by Grither at 1:33 PM on March 27, 2008

Response by poster: I, Safari.

Girlfriend, Seamonkey.
posted by popcassady at 1:35 PM on March 27, 2008

Response by poster: Router is a DG632, if that helps.
posted by popcassady at 1:36 PM on March 27, 2008

At the moment, it's not just you.
posted by katieinshoes at 2:48 PM on March 27, 2008 [1 favorite]

Response by poster: Now it says it's 'just me'.
posted by popcassady at 2:55 PM on March 27, 2008

If SBC is your ISP, then I have this guess: your own computer believes that it is actually sbc-services.com already. It looks for itself, finds no webserver on itself, fails. The ping works because it finds itself and answers, which is 'correct' behavior.

I have seen computers misconfigured this way (misinterpreting or guessing wrongly at the setup info an ISP provides) and the behavior is pretty much what you describe. Everything else works except the one domain that happens to be the ISP's domain.

You don't say Mac or PC. If Mac it's System Prefs > Network > Search Domains and System Prefs > Sharing. If a PC, it's My Computer > Computer Name and the "Join/Change Domain" button within. (That's Leopard and XP SP3, other versions will be a bit different.)

Instead of a ping, try a traceroute to sbc-services.com. How many hops? :)
posted by rokusan at 2:57 PM on March 27, 2008

Response by poster: No. SBC is not my ISP.

BTW, I'm using a Mac ...my girlfriend is using Linux.
posted by popcassady at 3:01 PM on March 27, 2008

Response by poster: I'm not sure how I'm meant to interpret these traceroute results. Are they in any way telling?

traceroute to sbc-services.com (78.47.122.133), 64 hops max, 40 byte packets
1 10.0.1.1 (10.0.1.1) 5.218 ms 6.491 ms 3.895 ms
2 * * *
3 anchor-hg-3-lo100.router.demon.net (194.159.161.34) 13.848 ms 15.218 ms 13.702 ms
4 anchor-access-4-s155.router.demon.net (194.159.161.161) 13.904 ms 14.376 ms 13.904 ms
5 anchor-inside-4-g6-0-4.router.demon.net (194.159.161.94) 14.383 ms 14.803 ms 14.148 ms
6 anchor-border-1-g4-0-0.router.demon.net (194.70.98.30) 16.385 ms 17.666 ms 14.736 ms
7 linx1.lon-2.uk.lambdanet.net (195.66.224.99) 15.017 ms 14.060 ms 33.748 ms
8 fra-1-pos313.de.lambdanet.net (82.197.136.25) 32.435 ms 32.074 ms 32.813 ms
9 fra-3-eth100.de.lambdanet.net (217.71.96.70) 33.142 ms 31.103 ms 32.077 ms
10 nue-2-eth100.de.lambdanet.net (217.71.96.74) 35.331 ms 35.876 ms 35.453 ms
11 lambdanet-gw.hetzner.de (213.239.242.214) 37.870 ms 38.684 ms 38.664 ms
12 hos-bb2.juniper3.rz4.hetzner.de (213.239.240.142) 40.272 ms 68.065 ms 41.116 ms
13 et.1.16.rs3k4.rz5.hetzner.de (213.239.244.35) 43.679 ms 60.736 ms 41.559 ms
14 * * *
15 * * *
16 * * *
17 * * *
..and all the same down to 64...
64 * * *

posted by popcassady at 4:53 PM on March 27, 2008

Maybe it's the same kind of MTU problem that I asked about and resolved in this MetaTalk thread?
posted by AmbroseChapel at 6:08 PM on March 27, 2008

No, the traceroute results are not telling, I'm afarid. I get to the same router and start showing *s until hop 64.

So, that DG632 is a modem _and_ a router, eh? Bummer. I had wanted you to narrow out the router as the possible problem, but without a DSL modem, you can't do that.

If you can login to the router, what is the MTU set to on the WAN page?

m
posted by tcv at 7:31 PM on March 27, 2008

Best answer: Router MTU was set to 1485. I tried it set to 1500 and 1470 but no positive results.

Ambrose, I tried using ping to determine the best MTU setting but the site won't ping from my computer.
posted by popcassady at 1:47 AM on March 28, 2008

DSL MTU is 1492. It should be set to that. Try that.
posted by tcv at 6:17 AM on March 28, 2008

Response by poster: 1492 seems to be the magic number... touch wood.

Thank you.
posted by popcassady at 12:55 PM on March 28, 2008

The MTU won't stop sites from loading. Clear your DNS cache?
posted by gjc at 4:26 PM on March 28, 2008

The more hops and packets needed to successfully receive a resource, the more chances for packet loss. I think one of the routers in between is not configured right, which accounts for the excessive amount of hops to get to the website. Running with a mis-set MTU would cause more packet loss and, perhaps, provide denial-of-service.

Or something.
posted by tcv at 8:54 PM on March 28, 2008

Response by poster: Well, five months later, the exact same problem has reared its ugly head for a second time.

Same site. Same traceroute results... Changing the MTU doesn't work this time.

Anything else I should try?
posted by popcassady at 2:04 PM on August 12, 2008

its not loading here either. sometimes a cigar is just a cigar.
posted by quonsar at 2:15 PM on August 12, 2008

Response by poster: Ah, that's interesting.

downforeveryoneorjustme.com was reporting it as okay about half an hour ago.
posted by popcassady at 2:23 PM on August 12, 2008

It's just plain old down. It's not you.
posted by majick at 2:51 PM on August 12, 2008

Response by poster: But now, downforeveryone... says it's up. Just not for me.
posted by popcassady at 3:24 PM on August 12, 2008

The traceroute you saw matches what I'm getting now, and the site loads for me. Various "script kiddie" idiocy over the years has led to traceroute being pretty unreliable: people realized they could 'ping flood' routers as a quick and easy Denial of Service attack, so now 'good' routers will completely ignore traceroute 'requests' if they're busy actually passing data, or just as a matter of policy.

traceroute to sbc-services.com (78.47.122.133), 64 hops max, 40 byte packets
 1  c-3-0-ubr01.mashpee.ma.boston.comcast.net (71.233.188.1)  8.429 ms  7.304 ms  5.889 ms
 2  ge-2-4-ur01.londonderry.nh.boston.comcast.net (68.86.236.97)  6.310 ms  6.323 ms  5.905 ms
 3  68.85.162.73 (68.85.162.73)  9.326 ms  8.783 ms  10.378 ms
 4  te-3-2-ar01.chartford.ct.hartford.comcast.net (68.86.90.62)  13.338 ms  11.837 ms  10.896 ms
 5  te-0-4-0-1-cr01.newyork.ny.ibone.comcast.net (68.86.90.61)  16.797 ms  14.297 ms  15.412 ms
 6  68.86.89.178 (68.86.89.178)  38.383 ms  14.351 ms  118.638 ms
 7  xe-0-0-0.fra20.ip.tiscali.net (89.149.186.65)  106.798 ms xe-3-2-0.fra20.ip.tiscali.net (89.149.187.49)  108.375 ms xe-0-0-0.fra20.ip.tiscali.net (89.149.186.65)  107.449 ms
 8  hetzner-gw.ip.tiscali.net (77.67.64.18)  115.883 ms  117.211 ms  116.55 ms
 9  hos-bb1.juniper1.rz4.hetzner.de (213.239.240.200)  117.459 ms  116.376 ms  116.33 ms
10  et.1.13.rs3k4.rz5.hetzner.de (213.239.244.163)  117.309 ms  117.690 ms  117.548 ms
11  * *^C

Another test might be to try to access it by IP, to determine whether it's a DNS problem or a network problem. http://78.47.122.133/ loads the same page for me. And also, see if your girlfriend gets the same thing?
posted by fogster at 5:01 PM on August 12, 2008

For what it's worth:

stephen@jellyfish:~$ tcptraceroute sbc-services.com 80
Selected device eth0, address 192.168.119.1, port 57080 for outgoing packets
Tracing the path to sbc-services.com (78.47.122.133) on TCP port 80 (www), 30 hops max

[snip routing out of Australia]

6 ge-6-20.car3.SanJose1.Level3.net (4.71.112.85) 195.010 ms 193.338 ms 196.191 ms
7 ae-34-89.car4.SanJose1.Level3.net (4.68.18.134) 198.969 ms 196.136 ms 191.038 ms
8 ge-1-1-1.sjc11.ip.tiscali.net (213.200.80.157) 232.978 ms 197.096 ms 194.726 ms
9 xe-0-0-0.fra20.ip.tiscali.net (89.149.186.65) 414.793 ms 353.094 ms 351.159 ms
10 hetzner-gw.ip.tiscali.net (77.67.64.18) 366.261 ms 369.022 ms 367.158 ms
11 hos-bb1.juniper1.rz4.hetzner.de (213.239.240.200) 358.613 ms 358.454 ms 375.299 ms
12 et.1.13.rs3k4.rz5.hetzner.de (213.239.244.163) 367.733 ms 368.289 ms 371.153 ms
13 static.133.122.47.78.clients.your-server.de (78.47.122.133) [open] 397.226 ms 400.552 ms 400.850 ms
stephen@jellyfish:~$
posted by flabdablet at 6:08 PM on August 12, 2008

Also:

stephen@jellyfish:~$ ping -c 4 213.239.244.163
PING 213.239.244.163 (213.239.244.163) 56(84) bytes of data.
64 bytes from 213.239.244.163: icmp_seq=1 ttl=239 time=362 ms
64 bytes from 213.239.244.163: icmp_seq=2 ttl=239 time=362 ms
64 bytes from 213.239.244.163: icmp_seq=3 ttl=239 time=362 ms
64 bytes from 213.239.244.163: icmp_seq=4 ttl=239 time=362 ms

--- 213.239.244.163 ping statistics ---
4 packets transmitted, 4 received, 0% packet loss, time 3000ms
rtt min/avg/max/mdev = 362.047/362.425/362.829/0.790 ms
stephen@jellyfish:~$ ping -c 4 et.1.13.rs3k4.rz5.hetzner.de
ping: unknown host et.1.13.rs3k4.rz5.hetzner.de
stephen@jellyfish:~$ dig et.1.13.rs3k4.rz5.hetzner.de

; <>> DiG 9.4.1-P1.1 <>> et.1.13.rs3k4.rz5.hetzner.de
;; global options: printcmd
;; Got answer:
;; ->>HEADER<>
posted by flabdablet at 6:22 PM on August 12, 2008

Response by poster: Nothing seems to be working.

So the problem is not my computer? Why is it just me?
posted by popcassady at 7:05 AM on August 16, 2008

What does tcptraceroute tell you?
posted by flabdablet at 5:24 PM on August 16, 2008

Response by poster: Well, I finally got round to installing darwinports.

I typed tcptraceroute sbc-services.com

and got:

Selected device en1, address 10.0.1.2, port 49788 for outgoing packets
Tracing the path to sbc-services.com (78.47.122.133) on TCP port 80 (http), 30 hops max
1 10.0.1.1 1.180 ms 0.892 ms 0.810 ms
2 * * *
3 * * *
4 * * *
5 * * *
6 * * *
7 * * *
8 * * *
9 * * *
10 * * *
11 * * *
12 * * *
13 * * *
14 sbc (78.47.122.133) [open] 40.865 ms * 39.385 ms

I have no idea what that means but -- I'm not sure if it's a coincidence or not -- the site loads fine now.

Did tcptraceroute fix it?
posted by popcassady at 1:19 AM on September 7, 2008

Are you really seeing "sbc" on hop 14, for 78.47.122.133? That address's reverse lookup is still "static.133.122.47.78.clients.your-server.de" for me.

Also, the fact that you're getting all those * * * lines strikes me as rather odd. Try running tcptraceroute against a bunch of other sites and see if the same thing happens. If so, I'd suspect your router of having been configured to drop all incoming ICMP messages (or at the very least, not handling NAT correctly for those messages) and I would not be at all surprised to find that this makes your networking act flaky. If you have the opportunity to try a different router, or try updated firmware in the one you've got, that may help.

Here's a quick rundown on how traceroute works, in case you need it (tcptraceroute also relies on being able to receive "TTL expired" ICMP messages from routers along the way to the target; the difference is that it provokes these by sending TCP SYN packets rather than ICMP ping requests).
posted by flabdablet at 6:40 AM on September 7, 2008

Response by poster: Well the old corpse is back. This evening, it started playing up again. I did the tcptraceroute thing again but this time it's all stars down to 'destination not reached'.

It seems to be a problem at my end. Router firmware is all up to date. I have no problems connecting to any other sites.

Arghh!!

What do I do? Do I spend pretty penny on another router? (FWIW, my current one is a Netgear DG632.) Do I give up?
posted by popcassady at 7:14 PM on October 11, 2008

Response by poster: ...and, at last, I've figured it out.


My IP address was blocked!
posted by popcassady at 12:52 PM on October 16, 2008